ec2006 instructions
A tool introduced
SPEC CPU 2006 benchmark is the standardization of the next generation of industry SPEC CPU benchmark suite. Key test system's processor, memory subsystem and compiler. The benchmark suite includes the standard SPECint and SPECfp benchmarks. Wherein the reference SPECint2006 comprises 12 different benchmarks and SPECfp2006年基准包含 19 different benchmark. SPEC designed this kit offers a relatively standard compute-intensive, high-performance inter-CPU hardware testing tool. There are several different ways in the SPEC CPU 2006 benchmark to measure the performance of the computer. One way is to measure the speed of the computer to perform one task; another way to measure certain capacity or rate. Note: Due to spec2006 support multiple types of operating systems. Following installation, testing, transplantation are introduced Unix-based and other Unix-like system, such as linux. Windows systems are not within this range.
1.spec2006 contrast brief description spec2000
August 24, 2006, Standard Performance Evaluation Corporation (SPEC) CPU2006 announced that will replace the CPU2000. Spec cpu benchmark is widely used in industry and academia. The new suite is much larger than cpu2000, and will cover the CPU, memory systems, compilers, especially C ++ compiler. Cpu2000 only one on C ++ benchmark, CPU2006 it contains seven related packages, including a 50W line of code c ++, Fortran and C language as before with good coverage.
2. Spec2000 and spec20006 kit specific information is as follows:
2.1 Spec int2000 12 integer-point test project
164.gzip C language written, very popular data compression program. Using the Lempel-Ziv coding (LZ77) as the compression algorithm. SPEC in addition to the read set gzip no other external file I / O operations, all compression and decompression are performed in the main memory, the CPU and the storage subsystem to facilitate testing.
Written 175.vpr C language, locating and routing circuitry in the FPGA program for computer aided circuit design.
176.gcc: C language, C language compiler optimization. The GCC-based GCC 2.7.2.2 version, generate machine code for the Motorola 88100 processor.
181.mcf: C language, combinatorial optimization / scheduling a single point of rotation (Single-depot vehicle scheduling), derived from a single station using a common rotation stream scheduling application.
186.crafty: C language, chess games. This is a design of high-performance machines on the 64-bit computer chess program may be used on a long long machine to represent the 32-bit data types. The main point of the program is a whole, a large number of logic operations. By the configuration, can generate different sets of search, can the whole point arithmetic processor, branch prediction, and the like for testing lines.
197.parser: C language, a word processing program. Connection parser (Link Grammar Parser) is connected to a grammar of English syntax analyzer based on. Earlier theoretical English grammar syntax connection. Given a sentence, the system give it gives a lexical structure that is between the two words are tagged connector. The analyzer has about 60,000 word dictionary table.
252.eon: C ++ language, computer visualization. eon is a probability-based ray tracing program, conference papers Kajiya 1986 ACM SIGGRAPH based. It certain amount of radiation transmitted to a 3D polyhedron in 3D, the intersection is calculated and polygons rays, and generates a new incident rays and calculating the intersection point. When the result of a camera can see in the picture the final calculation. Computational requirements and the basic computer graphics program in a conventional deterministic ray tracing procedure described is very similar, but it has less storage consistency requirements, because many rays randomly generated transverse to the 3D space in the same position, a completely different the area input:. eon a 150 × 150-pixel rendering of a chair in the room forefoot pictures.
253.perlbmk: C language, Perl language interpreter. This is a Perl v5.005_03 version of the cut version, delete a lot of operating system-related features, and adds some third-party modules, including MD5 v1.7, MHonArc v2.3.3, IO-stringy v1.205, MailTools v1.11, TimeData v1.08. Input: four perl script composed of an email-to-HTML converter, taking the longest time; spec CPU2000 to a comparison output of diff script, a script to find qualified digital iterative algorithm for testing a dummy a random number generator in a predetermined order script. Outputs corresponding to respective inputs respectively.
254.gap: C language, group theory, the interpreter used. It implements a library for language and do group theory calculations. GAP is (abbreviation Groups, Algorithms, and the Programming).
255.vortex: C language, a single user-oriented benchmark objects in a database transaction. SPEC the vortex from a subset of the full object-oriented database programs in VORTEx. Transactions are entered into the database and input is converted into a program.
256.bzip2: C language, compression program. Based on the Julian Seward bzip2 v0.1, the only difference is the SPEC bzip2 addition to the input, but no file I / O. All compression and decompression are carried out in memory. Input, output and gzip are similar.
300.twolf: C language, location and route simulation. twolf from TimberWolfSC global routing and positioning, which is the process used to create the microprocessor circuit layout of the printed product, and determining the global position of the group within the transistor. The problem is that a positioning arrangement, too time-consuming exhaustive, TimberWolfSC simulated annealing procedure is used to find a better solution as heuristic.
2.2 Spec int2000 12 integer-point test project
400.perlbench: C language, with 253.perlbmk spec2000 similar to that Perl V5.8.7 version of the cut version, delete the same operating system features a little something about the increase of third-party modules, such as: SpamAssassin V2. 61, Digest-MD5 v2.33, HTML-Parser v3.35, mHonArc v2.6.8, IO-stringy V1.205, MailTools V1.60, timeDate V1.16.400.perlbench uses three scripts or increased work load . 1. The main load is open spam filtering software SpamAssassin. Take the points system is the use of filtering, that will be given a score above a standard value when it is determined that SPAM (junk mail) according to the standard we set, rest not spam. But SpamAssassin software to do a large degree of modification, to avoid some of the I / 0 operation, and does not use Bayesian filters. 2. The email is more popular to HTML converter MHonArc 3. Use the tool kit of specdiff spec2006.
401.bzip2: C language, the same spec2000 256.bzip2 Similarly, based on the Julian Seward bzip2 SPEC V1.03 only difference is the addition bzip2 input, but no file I / O.. All compression and decompression are carried out in memory. Input, output and gzip are similar. Load comprises six parts: two small JPEG picture, a program, a tar wrapped several source files, an HTML file, mix files, including compressed files up high and not very compressible compressible files. We were tested using three different levels of compression to compress and decompress
403.gcc: C language, similar to the same spec2000 of 176.gcc. The test is based on gcc V3.2, machine code is generated AMDCPU. Compared spec2000 of 176.gcc, the test has more input files, so the pressure will be even greater test load of 9 from Group C code is compiled.
429.mcf: C language, also need libm library support, similar to the same spec2000 of 181.mcf, MCF is a large public transportation for the single station vehicle scheduling program. However, some of the input file changes made by the 32-bit into 64-bit, 64-bit system to be compatible. And increases the performance of cache hits and procedures. Compared spec2000 of 181.mcf speaking, 100M-190M previously occupied by the memory becomes 860M-1700M.
445.gobmk: C language, similar to the same 186.crafty spec2000. The difference is that here is the realization of the Go game. Compared spec2000 process more complicated.
456.hmmer: C language. HMMER is based on hidden Markov models (profile HMMs), for biological sequence analysis. Similar to the Timed HMMer Search.
458.sjeng: C language. Is based on a chess game Sjeng11.2, belong to the category of artificial intelligence.
462.libquantum: C language (C99). libquantum analog library is a quantum computer, used for quantum computer applications.
464.h264ref: C language. A video compression program, based on what H264AVC version 9.3, in addition to I / O and platform-dependent.
471.omnetpp: C ++ language. OMNeT ++, discrete event simulation. Including from about 8,000 computer 900 and switches / hubs, and a large variety of mixed to CSMA 1000Mb rate from 10Mb / CD Ethernet protocol simulation.
473 astar: C ++ language, 2D to achieve the three different versions of the A * pathfinding algorithm.
- xalancbmk: C ++ language, contain Xalan-C ++ v1.8 mod, Xerces-C ++ v2.5.0, implements XML documents / XSL table to convert the HTML document.
2.3 Spec fp2000 13 floating-point test project
168 wupwise: Fortran 77 language, lattice gauge theory and numerical simulation of Wilson fermions grid is calculated quark propagator. Wupwise.in input file contains several parameters of the problem created, SEED: providing a random number NITER: iterations KAPPA: parameters characterizing the matrix M = I - KAPPA * D. Output file wupwise.out, comprising NITER each iteration check value (EPS) and the final results comprising EPSILON features.
171 swim: Fortran 77 language, shallow water model meteorology.
172.mgrid: Fortran 77 language, multi-grid solver: 3D potential field
173.applu: Fortran 77 language, parabolic / Elliptic Partial Differential Equations
177.mesa: C language, 3D graphics library.
178.galgel: Fortran 90 language, computational fluid dynamics
179.art: C, Image Recognition / Neural Networks
183.equake: C language, seismic wave propagation simulation
187.facerec: Fortran 90 language, image processing: Face Recognition
189.lucas: Fortran 90 language, number theory / primality test
191.fma3d: Fortran 90 language, finite element simulation collision
200.sixtrack: Fortran 77 language, high-energy nuclear physics accelerator design
301.apsi: Fortran 77 language, meteorology distribution of pollutants
2.4 Spec fp2006 17 floating-point test project
410.bwaves: Fortran 77 language, hydrodynamics category. Simulated three-dimensional viscous flow instantaneous transonic shock wave is calculated
416.gamess: Fortran language, quantum chemistry category. Using GMAESS assembly, comprising cytosine, 3 SCF water and Cu2 + ions, self-consistent field ion triazole calculated.
433.milc: C language, the scope of quantum mechanics, using MILC assembly, comprising a four-dimensional sub quark and plastic SU (3) lattice gauge theory mold used to study QCD QCD.
434.zeusmp: Fortran 77 / REAL * 8 language, belonging to the scope of computational fluid dynamics, ZEUS-MP program, CFD and used in an ideal MHD, non-relativistic conditions, 434.zeusmp a uniform magnetic field simulation the 3D Blaster.
435.gromacs: C & Fortran language, biochemistry / molecular mechanics category. GROMACS is a suite of molecular mechanics calculations, but can also be used for non-biological systems, 435.gromacs simulate the changes in protein structure in a lysozyme solution of water and ions in various experimental means such as irradiation of X-rays NMR.
436.cactusADM: Fortran 90, ANSI C language, physics of general relativity. From Cactus BenchADM. Einstein 436.cactusADM evolution equation of curvature of space determined by the inner mass is solved by the second order nonlinear evolution equations Einstein 10 ADM 3 + 1 standard exploded partial differential equations.
437.leslie3d: Fortran 90 language, belongs to the category of fluid mechanics. LESlie3d is used to calculate the turbulent computational fluid dynamics program, 437.leslie3d a calculated time such as fuel into the combustion chamber of the stratified mixture fluid.
444.namd: C ++ language. Belonging biological / molecular visible. NAMD is a large biomolecular systems parallel program, it 444.namd simulate the apolipoprotein AI 92224 atoms
447.dealII: C ++ language. Finite element analysis, w / Boost lib, deal.II lib, deal.II is located in the adaptive finite element and error estimates of C ++ libraries, 447.dealII coefficient of Helmholtz equation is solved, it uses the generating an adaptive estimation method based on the binary grid best weighted error, the three-dimensional equations have solutions.
450.soplex: C ++ language. Procedure using Plex v1.2.1, Plex simplex algorithm solving linear equations
453.povray: ISO C ++ language. POV-Ray POV-Ray is a ray tracing rendering software, 453.povray anti-aliasing rendering a chess board image of 1280x1024
454.calculix: Fortran 90 & C language, structural mechanics category. CalculiX is a finite element analysis software for the three linear and nonlinear structural mechanics, 454.calculix The stress and deformation of a disk rotating at high speed under the effect of compression of the centrifugal force
459.GemsFDTD: Fortran 90 language. It belongs to the category of electromagnetism. 459.GemsFDTD using the FDTD (Finite Difference Time Domain) method for solving three-dimensional Maxwell equations in the time domain, calculates an ideal conductor RCS
465.tonto: Fortran 95 implementation. Quantum Chemical visible Tonto is an object-oriented quantum chemistry package, 465.tonto for quantum computing crystallography, which is based on an experimental X-ray diffraction data conform to the constraints of molecular Hartree-Fock wave functions
470.lbm: ANSI C language. 470.lbm using the LBM (Lattice Boltzmann Method) simulated non-compressed fluid, which simulates two cases: Similar shear piston driving fluid and driven fluid conduit, the test comprising the steps 3000
481.wrf: Fortran 90 & C language. WRF v2.0.2 481.wrf based WRF (Weather Research and Forecastin) model, NCAR data were calculated, including the UTC 2001.06.11 data to UTC 2001.06.12 data in a three-hour intervals
482.sphinx3: C language. Sphinx-3 A speech recognition software.
Second, the installation tool
1. Standard Installation
$mount -t iso9660 -o ro,exec /dev/cdrom /mnt
$cd /mnt
$./install.sh
You can specify the installation directory, such as:
$./install.sh -d /home/cpu2006
You may encounter during installation the following tips:
Installing FROM /Volumes/SPEC_CPU2006
Installing TO /home/cpu2006
Is this correct? (Please enter 'yes' or 'no')
yes
2. Perform a test done by installing software
First prepare the relevant test file, such as: cpu2006-20141030.tar. unzip files:
$ Tar xf cpu2006-20141030.tar
Into the installation directory:
$ Cd cpu2006v99 / (actual path)
Import test tool:
$. shrc
Note: When the input, and at least one space between dot SHRC, the import start command program SPEC CPU2006. Modify the absolute path of the test file:
$bin/relocate
显示“Top of SPEC benchmark tree is '/cpu2006v99'”
3. Run the test
According to needs, preferences, began to run the test script: For example: runspec -c linux-mipsel-gcc -i ref -n 3 all command parameters:
-c: reading the test profile, linux-mipsel-gcc config directory located SPEC CPU2000 installation directory, the file can be modified based on the actual information. It is not recommended to modify the compiler options.
-i: test input data set size, ref represents the maximum test data set, representing the minimum test Test data set;
-n: item number of each test run, if necessary SPEC CPU2006 automatically calculate test scores, required for the operation is equal to a specified number or greater than 3, i.e., n-> = 3;
All: all test runs the SPEC CPU2006 benchmark program items.
The following message appears, indicating that the test has started. "
runspec v4283 - Copyright 1999-2006 Standard Performance Evaluation Corporation
Using 'mipsel-linux' tools
Reading MANIFEST... 17899 files
Loading runspec modules...............
Locating benchmarks...found 31 benchmarks in 13 benchsets.
Locating output formats: ASCII, config, CSV, flags, html (DISABLED), mail, PDF, PostScript, raw, Screen, Submission Check
Reading config file '/cpu2006v99/config/linux-mipsel-gcc.cfg'
Benchmarks selected: 400.perlbench, 401.bzip2, 403.gcc, 429.mcf, 445.gobmk, 456.hmmer, 458.sjeng, 462.libquantum, 464.h264ref, 471.omnetpp, 473.astar, 483.xalancbmk, 999.specrand, 410.bwaves, 416.gamess, 433.milc, 434.zeusmp, 435.gromacs, 436.cactusADM, 437.leslie3d, 444.namd, 447.dealII, 450.soplex, 453.povray, 454.calculix, 459.GemsFDTD, 465.tonto, 470.lbm, 481.wrf, 482.sphinx3, 998.specrand
Compiling Binaries
up to date 400.perlbench ref base linux-mipsel-gcc default
Up to date 401.bzip2 ref base linux-mipsel-gcc default
Up to date 403.gcc ref base linux-mipsel-gcc default
Up to date 429.mcf ref base linux-mipsel-gcc default
……
This basic set of test configurations and spec2000 similar, except that, spec2006 before the test, each test component will establish a recompile. So before the test, make sure the system at this time contains the following compiler: g ++, gcc, gfortran compiler. If continuous testing, it is not necessary to recompile, you can add the parameter "-nobuild" general description attached the config file: // italics text is explanation
#####################################################################
ignore_errors = yes
tune = base
ext = loongson.3a2000
output_format = asc, Screen, html
reportable = 0
teeout = yes
teerunout = yes
hw_avail = Feb-2015
sw_avail = Feb-2015
license_num =
company_name = Loongson
prepared_by = Loongson
tester_name = isoft
test_date = July-2016
verbose = 9
default=default=default=default:
#####################################################################
#
# Compiler selection
#
#####################################################################
makeflags = -j 4
CC = mips64el-loongson-linux-gcc
CXX = mips64el-loongson-linux-g++
FC = mips64el-loongson-linux-gfortran
## HW config
hw_cpu_name = Loongson
hw_model = 3A2000
hw_cpu_mhz = 1000
hw_fpu = Integrated
hw_nchips = 1
hw_ncores = 4
hw_ncoresperchip= 4
hw_nthreads = 4
hw_ncpuorder = 1
hw_parallel = yes
hw_pcache = 64KB I + 64KB D
hw_scache = 4096KB(I + D)
hw_memory = 2 x 4GB DDR1066
hw_vendor = Loongson
## SW config
sw_os = iSoft
sw_file = ext4
sw_state = runlevel 3
sw_compiler = gcc , g++ & gfortran 4.8
#####################################################################
# Optimization
#####################################################################
## Base is low opt
default=default=default=default:
# For this data model, -m32 needs to be here, as one of the lines below
OPTIMIZE = -static -march=loongson3a
PORTABILITY =
COPTIMIZE = -flto -Ofast
CXXOPTIMIZE = -flto -Ofast
FOPTIMIZE = -flto -Ofast
notes0100= C/C++ base flags: -flto -Ofast
notes0110= Fortran base flags: -flto -Ofast
#####################################################################
# 32/64 bit Portability Flags - all
#####################################################################
default=base=default=default:
notes25=
default=peak=default=default:
notes25=
#####################################################################
# Portability Flags - INT
#####################################################################
400.perlbench=default=default=default:
notes35 = 400.perlbench: -DSPEC_CPU_LINUX_IA32
CPORTABILITY= -DSPEC_CPU_LINUX_IA32 -fno-strict-aliasing
#notes35 = 400.perlbench: -DSPEC_CPU_LINUX
CPORTABILITY = -DSPEC_CPU_LINUX -DSPEC_CPU_LITTLEENDIAN #
#COPTIMIZE = -O2
......
compiling information on other specific components no longer them out, which can keep the default.
Test methods and more types see below.
4. Test results are collected and analyzed
After the test result in the catalog, view the test results, test results are generally the type of test and test order to distinguish, and will be divided according to the whole point and floating point. If the test is carried out in accordance with official reportable mode, it will be the official test results of a web version, see the following figure: unofficial results can go asc or txt document viewing format such as the following:
Base Base Base Peak Peak Peak
Benchmarks Ref. Run Time Ratio Ref. Run Time Ratio
-------------- ------ --------- --------- ------ --------- ---------
400.perlbench 9770 430 22.7 S 9770 339 28.8 S
400.perlbench 9770 429 22.8 * 9770 338 28.9 S
400.perlbench 9770 429 22.8 S 9770 339 28.8 *
401.bzip2 9650 559 17.3 * 9650 552 17.5 *
401.bzip2 9650 559 17.3 S 9650 553 17.5 S
401.bzip2 9650 560 17.2 S 9650 552 17.5 S
403.gcc 8050 374 21.5 S 8050 337 23.9 S
403.gcc 8050 367 22.0 * 8050 337 23.9 S
403.gcc 8050 366 22.0 S 8050 337 23.9 *
429.mcf 9120 187 48.9 S 9120 187 48.9 S
429.mcf 9120 188 48.6 * 9120 188 48.6 *
429.mcf 9120 211 43.2 S 9120 211 43.2 S
445.gobmk 10490 559 18.8 S 10490 542 19.4 S
445.gobmk 10490 559 18.8 S 10490 542 19.4 *
445.gobmk 10490 559 18.8 * 10490 541 19.4 S
456.hmmer 9330 213 43.8 S 9330 213 43.7 *
456.hmmer 9330 247 37.8 S 9330 214 43.6 S
456.hmmer 9330 214 43.5 * 9330 213 43.7 S
458.sjeng 12100 616 19.6 * 12100 574 21.1 *
458.sjeng 12100 617 19.6 S 12100 575 21.0 S
458.sjeng 12100 585 20.7 S 12100 574 21.1 S
462.libquantum 20720 14.2 1460 * 20720 14.2 1460 *
462.libquantum 20720 14.4 1440 S 20720 14.4 1440 S
462.libquantum 20720 13.5 1540 S 20720 13.5 1540 S
464.h264ref 22130 578 38.3 * 22130 527 42.0 S
464.h264ref 22130 579 38.2 S 22130 526 42.0 *
464.h264ref 22130 578 38.3 S 22130 526 42.0 S
471.omnetpp 6250 349 17.9 S 6250 288 21.7 *
471.omnetpp 6250 348 17.9 * 6250 288 21.7 S
471.omnetpp 6250 348 18.0 S 6250 289 21.7 S
473.astar 7020 313 22.4 S 7020 313 22.4 S
473.astar 7020 311 22.6 S 7020 311 22.6 S
473.astar 7020 311 22.6 * 7020 311 22.6 *
483.xalancbmk 6900 174 39.7 * 6900 170 40.7 S
483.xalancbmk 6900 174 39.6 S 6900 170 40.5 S
483.xalancbmk 6900 174 39.7 S 6900 170 40.6 *
==============================================================================
400.perlbench 9770 429 22.8 * 9770 339 28.8 *
401.bzip2 9650 559 17.3 * 9650 552 17.5 *
403.gcc 8050 367 22.0 * 8050 337 23.9 *
429.mcf 9120 188 48.6 * 9120 188 48.6 *
445.gobmk 10490 559 18.8 * 10490 542 19.4 *
456.hmmer 9330 214 43.5 * 9330 213 43.7 *
458.sjeng 12100 616 19.6 * 12100 574 21.1 *
462.libquantum 20720 14.2 1460 * 20720 14.2 1460 *
464.h264ref 22130 578 38.3 * 22130 526 42.0 *
471.omnetpp 6250 348 17.9 * 6250 288 21.7 *
473.astar 7020 311 22.6 * 7020 311 22.6 *
483.xalancbmk 6900 174 39.7 * 6900 170 40.6 *
SPECint(R)_base2006 36.7
SPECint2006 39.1
Base Ratio test value, the test is generally run three times to take the median value. The overall result is the geometric mean of all results.
5 illustrates a common test command combination SPEC CPU2006
1. runspec -c test.cfg -i test -I all test based on a minimum set of data quickly perform all the tests, the testing process, if an error occurs with the embodiment, a skip error cases, to continue other use cases. 2. runspec -c test.cfg -i ref -n 3 -I all the full implementation of the test data set based on the maximum all the tests, for testing single-core CPU, the testing process if an error occurs use cases, use cases skip the error, continue perform other use cases. 3. runspec -c test.cfg -r ref -n 3 fp based on the maximum test data set, the test run only fp 4. runspec -c test.cfg -r ref -n 3 int test data set based on a maximum run only int test 5.runspec -c 473.astar single test test.cfg -i ref 473.astar perform only test data set based on a maximum. 6 runspec -c test.cfg -i ref --rate 4 int it is based on the maximum data rate test test set, test run score 4 threads
6. Other supplementary command
--action build create a test project
--action onlyrun just run, do not confirm whether the test results correctly
--action report standard test
test project clean clean-up has been established --action
Mode a lot, no longer introduced one by one.
--check_version check the version update
--copies -c number 1 to be the default RUN SPECrate
--help display help document
--ignore_errors ignore the error
ations number -n number of tests, a reportable test, must be 3 times
--iter-loose No reportable test set
--output_format output file format default html text may be selected in various ways all cfg test csv html mail pdf ps printer raw screen text and the like.
--rate -r be SPECspeed test you can specify the number of copies run.
--rebuild recompile the test file
--tune tuning can select the default base base, peak, or all, Report First base, followed by peak.
These are more commonly used commands, some of the less frequently used commands not described, may be understood by reference below.
-a Same as --action
--action action Do: build|buildsetup|clean|clobber|configpp| onlyrun|realclean|report|run|runsetup|scrub| setup|trash|validate
--basepeak Copy base results to peak (use with --rawformat)
--nobuild Do not attempt to build binaries
-c Same as --config
-C Same as --copies
--check_version Check whether an updated version of CPU2006 is available
--comment "text" Add a comment to the log and the stored configfile.
--config file Set config file for runspec to use
--copies Set the number of copies for a SPECrate run
-D Same as --rebuild
-d Same as --deletework
--debug Same as --verbose
--define SYMBOL[=VALUE] Define a config preprocessor macro
--delay secs Add delay before and after benchmark invocation
--deletework Force work directories to be rebuilt
--dryrun Same as --fake
--dry-run Same as --fake
-e Same as --extension
--ext Same as --extension
--extension ext[,ext...] Set the extensions
-F Same as --flagsurl
--fake Show what commands would be executed.
--fakereport Generate a report without compiling codes or doing a run.
--fakereportable Generate a fake report as if "--reportable" were set.
--[no]feedback Control whether builds use feedback directed optimization
--flagupdate Same as --update
--flagsupdate Same as --update
--flagsurl url Location (url or filespec) where to find your flags file
--getflags Same as --update
--graph_auto Let the tools pick minimum and maximum for the graph
--graph_min N Set the minimum for the graph
--graph_max N Set the maximum for the graph
-h Same as --help
--help Print usage message
--http_proxy Specify the proxy for internet access
--http_timeout Timeout when attempting http access
-I Same as --ignore_errors
-i Same as --size
--ignore_errors Continue with benchmark runs even if some fail
--ignoreerror Same as --ignore_errors
......
7. manually compile and cross-platform method (for reference only)
Spec2006 included in the test suite is essentially independent of the platform, so in theory can be directly cross-platform testing. Spec2006 In order to achieve accurate and reduce dependence on the platform of the test results, produced its own toolkit. These include: specdiff (examines results to see if the correct answer was obtained), specinvoke (invokes benchmarks for CPU2006), specmake (GNU make, Calling it "specmake" avoids possible conflicts with versions of make that may already be on your system. SPEC requires that published results use the versions of the tools that it supplies, so that if SPEC applies patches or extensions from time to time, all users run with a consistent tool set. Similar considerations apply to other tools in this list.), specxz (The xz compression program), specmd5sum (md5sum from GNU coreutils, with enhancements), spectar (GNU tar), specperl (Perl). Various Perl modules such as Algorithm :: Diff, GD, HTML :: Parser, PDF :: API2, URI, XML :: SAX, etc .. Compilation Tools: C99 compiler C ++ compiler if initial installation package, the package need to extract the source
# xz -dc install_archives/cpu2006.tar.xz | tar -xf - tools/src
Command to establish Toolkit as follows:
AIX (32-bit):
CC=xlc CFLAGS='-O -qarch=ppc -D_LARGE_FILES -D_ILS_MACROS' PERLFLAGS="-Doptimize='-O' -Accflags='-D_LARGE_FILES'" OBJECT_MODE=32 ./buildtools
AIX (64-bit):
CC=xlc CFLAGS='-O3 -qarch=ppc64 -q64 -D_ILS_MACROS' PERLFLAGS="-Doptimize='-O' -Aoptimize='-qarch=ppc64 -q64'" OBJECT_MODE=64 ./buildtools
HPUX:
CC="/opt/ansic/bin/cc" CFLAGS="-Ae +O2 +z -D_FILE_OFFSET_BITS=64 -AC99" XZCFLAGS=-AC99 PERLFLAGS="-Doptimize=+O1 -Dlibpth=/usr/lib/hpux32 -Dldflags=-lm -Dcppstdin=cc -Acppstdin=-E -Acppstdin=-Aa -Acppstdin=-D__STDC_EXT__ -Acppstdin=-AC89" ./buildtools
IRIX:
ulimit -s 128000; CFLAGS="-c99 -D_FILE_OFFSET_BITS=64" MD5CFLAGS="-O2" PERLFLAGS="-Doptimize=-O3 -Uloclib -Dlibpth=/usr/lib32" XZFLAGS="--disable-assembler --disable-shared" XZCFLAGS="-O1" ./buildtools
After the completion of the establishment, confirming the success of the method:
cd $SPEC (Unix)
. shrc
或 runspec -V
或specmake -v
specxz -h
specperl -v
specdiff –h
runspec -h
specinvoke -h
After the completion of the translation tools, the need for tools to package
packagetools <archname>
If problems arise automated compiler, you can manually compile each test tool. Compiled into the method corresponding kit, manually compiled.